Winners 2025
Award Operations Research/Statistics
Dr. Nils Koster: Simplifying Machine Learning Models: Balancing Complexity and Performance
About the Dissertation: This dissertation explores how machine learning models can be made simpler while preserving the original predictive performance. It is motivated by the idea that simpler models, or simpler representations of complex models, are often easier to apply and to interpret. In this sense, the work can be seen as a step toward models that may eventually become easier to train and less resource-demanding in practice. The thesis develops methods to simplify neural networks and random forests, and applies a straightforward kernel-based method to tandem mass spectrum prediction. Lastly, a novel, interactive course format is presented that teaches probabilistic forecasting approaches that work well in practice - most often, the simple ones.
One-minute video about the dissertation on YouTube
Award Business Administration
Dr. Saskia Jacob: Adaption durch Sprache und Stimme: Dynamische Analyse von Mitarbeiter-Kunden-Interaktionen
About the Dissertation: The dissertation examines how linguistic and vocal adaptation in employee-customer interactions influences trust, emotions, and sales performance.
A key limitation of classical approaches is their limited ability to capture the temporal dynamics and complexity of social interactions. Using machine learning methods, this study analyzes and quantifies dynamic interaction processes across the course of conversations. The empirical part comprises three studies on emotional contagion and linguistic mirroring, based on both experimental and real-world sales data.
The dissertation demonstrates how these processes can be captured at a finer level of granularity, thereby providing methodological and theoretical contributions to the analysis of employee-customer interactions.
Award Information Systems
Dr. Philipp Spitzer: The Impact of Explainable AI on Knowledge in Human-AI Interactions
About the Dissertation: While artificial intelligence (AI) systems achieve high performance levels, a pivotal challenge lies in their effective integration into human workflows. Current research focuses on developing AI systems that assist humans by providing advice to augment their capabilities. However, their “black-box” nature often obscures the underlying reasoning processes. Explainable AI (XAI) has emerged to enhance the interpretability and usability. However, with the integration of XAI into human workflows, its impact on knowledge and the capability to independently perform tasks is poorly understood. By investigating interaction settings, this dissertation builds a theoretical understanding of the impact of XAI on knowledge development and sets out guidelines to design effective human-AI collaborations.
Award Informatics
Dr. Marc Zofka: Concept and Development of Virtually Enhanced Test Areas for Testing Highly Automated Driving in Urban Environments
About the Dissertation: This dissertation researches the design of a test system based on a digital twin, with smart roadside infrastructure for proving grounds and test areas in order to generate urban test scenarios while ensuring criterias such as observability and controllability. Contributions include the capability model for proving grounds and test areas, the immersive integration of vulnerable road users, the augmentation of sensor and environmental models along various injection points, and the concept of a distributed test system. The empirical evaluation with test vehicles at the „Test Area Autonomous Driving Baden-Württemberg“ enables the assessment of the researched methods.
Winners 2024
Award Operations Research
Dr. Christian Füllner: On approximating non-convex value functions in stochastic dual dynamic programming and related decomposition methods
About the Dissertation: "Stochastic Dual Dynamic Programming“ (SDDP) is considered the state-of-the-art algorithm for solving multi-stage stochastic linear optimization problems. One of its core ideas is the decomposition of the original problem into subproblems that are coupled via convex value functions, which are then iteratively approximated by means of cutting planes (“cuts”). However, many practical applications require integer conditions or nonlinear functions in the modeling, under which the convexity of the value functions is no longer guaranteed. In this dissertation, extensions of SDDP and its cuts to various cases of non-convex value functions are presented, analyzed with regard to their mathematical properties and evaluated in applications to case studies (e.g. hydrothermal scheduling).
Award Informatics
Dr. Meike Ullrich: Competency-oriented E-Assessment for Graphical Modeling in Higher Education
About the Dissertation: Graphical modeling plays an important role in computer science, e.g., in software development and business process management. Current university teaching hardly considers competency-oriented approaches, and technical support remains insufficient. In this dissertation, a competency model for learning objectives in graphical modeling and cross-disciplinary skills, an exercise catalog, assessment schemes, and feedback recommendations are developed. Additionally, an e-assessment platform is implemented to support various modeling languages, custom assessment schemes, and automated feedback for student models. It enables self-directed learning for students and eases digital exam grading for instructors.
Award Business Administration
Dr. Fabian Eska: An Asset Pricing Perspective on the Design of Cryptocurrencies)
The dissertation examines how specific design features - such as the consensus mechanism, the reward system or the degree of anonymity - influence the value, volatility and risk of cryptocurrencies.
Based on a specially developed taxonomy, extensive empirical analyses show that technical decisions in the code and decisions relating to network structures have measurable economic consequences.
In particular, the focus is on the comparison between proof-of-work and proof-of-stake - with regard to network security and climate risks, theoretically substantiated by model-based analyses.
The work combines classical financial market theory with modern crypto technology and provides valuable insights for investors, developers and regulators.
Award Economics/ Statistics
Dr. Claudio Kretz: Essays on Individual and Collective Decision Making
This thesis contains four essays on topics in individual and collective decision making. First, I develop an axiomatic model of decision makers who use plans to exert self-control when expecting preferences to change. Second, I reformulate the Arrow (1951) problem of independent and non-dictatorial aggregation in the framework of choice functions satisfying more general rationality conditions. Third, I characterize when rights granted to different subgroups are consistent to the effect that some aggregation procedure respects these rights for every possible scenario of how they might be exercised. Fourth, I compare two simple approaches for aggregating judgments in terms of their epistemic properties when voters are known to be correlated.
Award Interdisciplinary work
Dr. Max Kleinebrahm: Future Residential Energy System Design
This dissertation explores how emerging technologies reshape energy demand in the residential building sector. It introduces an AI-based approach to simulate realistic occupant behavior and model future electricity, heat, and mobility needs. A bottom-up optimization framework is applied to assess the potential of 41 million single-family homes in Europe for off-grid self-sufficiency. Additionally, a municipal planning model is extended to support cities in transforming their energy systems in line with national climate goals. The work integrates technical, economic, and behavioral perspectives to guide the design of climate-neutral energy systems.
Winners 2022
Award Business Administration
Dr. Michael Reichenbacher: Managing Liquidity Risks in Bond Markets
About the Dissertation: For investors in bond markets, the implementation and costs of many trading strategies depend on the underlying liquidity. In particular, both the initial liquidity of a bond and the liquidity available at the expected time of sale are of crucial importance. Against this background, the dissertation deals with the management of liquidity risks on bond markets. The focus here is on the quantification of these risks through precise measurement and prediction of liquidity, as well as the development of statistical tests to evaluate the prediction models.
Award Informatics
Dr. Ivo Benke: Emotional Competence Development Systems for Remote Work Collaboration
About the Dissertation: Distributed work enabled by digital collaboration technologies such as video meetings and group chats allows people to work together across physical boundaries. However, the use of these technologies makes social and emotional communication more difficult. It is therefore important to develop a high level of emotional competence - the ability to perceive, understand and regulate emotions. This dissertation deals with the design of intelligent systems for the development of emotional competence. Specifically, the dissertation develops retrospective analysis systems for video meetings and conversational agents for group chats that support emotional competence development and investigates their impact on feelings of autonomy and trust during use.
Award Interdisciplinary work
Dr. Christoph Fraunholz: Market Design for the Transition to Renewable Electricity Systems
About the Dissertation: The dissertation examines the role of market design in the transition to a future electricity system that ensures a sustainable, affordable and reliable electricity supply. For this purpose, an agent-based electricity market model is first methodically extended and then used to analyze the diffusion of different technologies under different regulatory frameworks. The focus is on three central aspects of the European electricity market: capacity mechanisms, the interaction between the day-ahead market and congestion management, and the regulatory framework for household battery storage. The results confirm that policymakers and regulators can significantly influence the composition, location and operation of technologies by shaping the regulatory framework.
Award Operations Research
Dr. Christoph Neumann: Inner Parallel Sets in Mixed-Integer Optimization
About the Dissertation: This paper contains a comprehensive study of inner parallel sets in mixed-integer optimization. Inner parallel sets are a new idea in this context and offer a way to reduce the difficulties imposed by integer constraints by guaranteeing the admissibility of roundings of their (continuous) elements. In order to be able to use inner parallel sets algorithmically, various modifications are necessary, such as their enlargement and inner and outer approximation. Such ideas are introduced and investigated in this thesis.
The potential of the resulting algorithms is demonstrated in numerical investigations on standard libraries for mixed-integer minimization problems.
Winners 2022
Award Business Administration
Dr. Marcel Müller: Asset Pricing Bricks
About the Dissertation: The dissertation deals with risks that have an influence on security prices. Specifically, three central research questions are examined: 1. how does the market's assumption that a government default can be excused influence the price of government bonds? -->Excusable defaults are priced in with significantly higher discounts. 2. to what extent do fund managers of investment funds exploit their knowledge advantage over investors? -->Fund managers try to increase the returns of their funds by investing in hidden risks. 3. is distorted media coverage a systematic risk factor for equities? -->If distorted media coverage is integrated as an additional factor in a price model, the quality of the model improves significantly.
Current Job Position: Freelance Consultant
Award Operations Research
Dr. Katharina Glock: Emergency rapid mapping with drones. Models and solution approaches for offline and online mission planning
About the Dissertation: The use of unmanned aerial vehicles (UAVs) makes it possible, for example after major fires, to provide emergency services with an initial picture of the situation in a short space of time. In this context, the dissertation deals with the planning of UAV missions that maximize the information gain in emergency operations. For this purpose, both offline heuristics, which determine missions before take-off, and online strategies, which update plans during the flight of the drones based on learned information, are considered and evaluated. The overall goal is to design efficient models and procedures that use information about the spatial correlation in the observed area to determine solutions of high predictive quality in time-critical situations.
Current Job Position: Senior Expert at FZI Research Center for Information Technology.
V
Award Economics/Statistics
Dr. Lars Herberholz: On the Mechanics Behind Academic Progress
About the Dissertation: Both the generation of new knowledge and the training of a highly qualified workforce play a key role in modern societies. However, recent studies show that new ideas are becoming increasingly rare and that innovation potential is often dwindling. Against this background, the dissertation is dedicated to the mechanisms underlying academic progress. Specifically, the thesis examines the forms in which spillover effects occur in research networks, which types of organization enable the efficient production and transfer of knowledge and whether a university funding programme set up by the Federal Ministry of Education and Research was designed to meet the objectives.
Current Job Position: Postdoc
VAward Business Administration
Dr. Michael Hofmann: Frictions, Intermediaries, and the Option Market
About the Dissertation: Michael Hofmann's dissertation entitled "Frictions, Intermediaries, and the Option Market" deals with the interplay of frictions (i.e. imperfections in the financial markets) and price formation on the stock option market. The focus is on the valuation implications of limited capital availability, the specific reaction of the option market to mispricing in the equity market, and the analysis of option price-implicit information about frictions in the option market and beyond.
Current Job Position: Senior Risk Methodologist at LBBW
Award Informatics
Dr. Peter Mayer: Secure and Usable User Authentication
About the Dissertation: Attacks on passwords are one of the biggest IT security threats for companies and private individuals. Progress in combating these attacks is therefore of importance to society as a whole. This dissertation describes progress in all three areas relevant in this context. Specifically, it provides (a) a process for the systematic development of demonstrably effective awareness and training materials to better protect users, (b) the first investigation of the risk of observational attacks when entering passwords with limited input devices as pioneering research with respect to smart devices, and (c) a method for securely and efficiently storing passwords in forward-looking alternatives to text passwords.
Current Job Position: Postdoc
Award Operations Research
Dr. Peter Kirst: Disjunctive Aspects in Generalized Semi-infinite Programming
About the Dissertation: The dissertation deals with the relationship between generalized semi-infinite problems (GSIP) and disjunctive problems (DP).
theory and solution methods for both classes of problems are investigated. A new way of representing DPs as GSIPs is presented. Known reformulation techniques for this latter problem class then lead to a new solution approach for disjunctive problems. In addition to this local solution procedure, new methods for the global optimization of the two problem classes are also derived.
The diverse application possibilities of DPs and GSIPs are illustrated using three examples from the timber industry, such as the calculation of optimal trusses and waste minimization.
Current Job Postion: Researcher at "Wageningen University & Research" , Netherlands
Award Interdisciplinary Work
Dr. Felix Hübner: Planning and modeling of large-scale projects under consideration of uncertainties using the example of the decommissioning of nuclear facilities
About the Dissertation: Large-scale projects such as the dismantling of nuclear facilities often have cost overruns and longer than planned execution times, partly due to inadequate planning. As part of this dissertation, a planning tool for large-scale projects was developed that takes into account the requirements of nuclear decommissioning projects as an example. The aim of the planning is to identify a plan at the operational level that minimizes costs and takes uncertainties into account.
Current Job Position: Strategy and Corporate Development Officer at VSE AG
Award Business Administration
Dr. Verena Rieger: Culture and innovation: Empirical studies at country, organizational and team levels
About the Dissertation: This work uses a series of empirical studies to shed light on the role of culture as a driver of innovation at country, organizational and team level. First, a value system for the conceptualization of an innovation-promoting organizational culture is developed as part of a literature synthesis. Subsequently, a laboratory experiment and a field study are used to provide evidence of the causal relationship between culture and innovation performance at team level. Finally, the relevance of national culture for the innovation performance of organizations is demonstrated on the basis of a large-scale survey of companies in 28 countries.
Current Job Position: Research assistant and habilitation candidate at the Chair of Business Administration, in particular Management at Heinrich Heine University Düsseldorf
Award Informatics
Dr. Patrick Philip: Decision-Making with Multi-Step Expert Advice on the Web
About the Dissertation: The thesis deals with solving multi-step tasks by using advice from experts, which are algorithms to solve individual steps of such tasks. We contribute with methods for maximizing the number of correct task solutions by selecting and combining experts for individual task instances and methods for automating the process of solving tasks on the Web, where experts are available as Web services.
Current Job Position: Senior Expert "Applied Reinforcement Learning" at FZI Research Center for Information Technology
Award Operations Research
Dr. Maximilian Coblenz: Advances in Dependence Modeling: Multivariate Quantiles, Copula Level Curve Lengths, and Non-Simplified Vine Copulas
About the Dissertation: Copulas make it possible to separate the dependency structure of a multivariate random variable from its univariate marginal distributions. Therefore, they are a powerful tool to study, model and interpret the dependence of multivariate data. The thesis contributes to
dependency modeling by addressing theoretical and application-oriented aspects of copulas,
which are of increasing importance in fields such as hydrology and risk management.
Current Job Position: Senior Data Scientist at 1&1 Telecommunication SE
Award Economics/ Statistics
Dr. Chong Liang: High Dimensional Time Series - New Techniques and Applications
About the Dissertation:
The past decade witnessed the rapid development of high dimensional statistics in deterministic design. High dimensional time series analysis, due to the time dependency, still faces several theoretical challenges. Among the time series models, the Vector Error Correction Model (VECM) is especially complicated because of the non-stationary components. The classical estimation strategies (e.g. Johansen’s approach) fail to provide consistent estimates for dimensions larger than three. Moreover, it is impossible to apply existing statistical methods to determine VECM in high dimensions. This dissertation aims at providing feasible regularized methods, which can determine and estimate high dimensional VECM with robust statistical properties.
Current Job Position: Quant Analyst at Credit Suisse
Award Interdisciplinary Work
Dr. Jan Ihrig: Spatial disaggregation in transport modelling
About the Dissertation: In European transport policy consulting, computer models are applied to investigate the effectiveness as well as the welfare impacts of potential policy measures. However, the travel zones used in recent models operating at European scale are often too large, mainly due to complexity and data availability. These models provide only limited insights into regional traffic flows.
In this thesis, the HIPAT model is introduced, which is based on disaggregated travel zones, and thus facilitates the modelling of European traffic flows, including long-distance, regional and short-distance passenger trips with a single, consistent transport model. This enables European policy makers to assess also regional welfare impacts when prioritising, for instance, investments in the trans-European transport network.
Current Job Position: Senior Expert Forecasting, Optimization, Simulation at Deutsche Bahn AG